Innovative Applications and Challenges of Isothermal Amplification Technology in miRNA Detection

Abstract

A class of endogenous non-coding RNAs with a length of roughly 18–25 nucleotides is known as microRNAs (miRNAs). They have been established as the best biomarkers for early cancer diagnosis, molecular subtyping, and prognostic evaluation. They also play important roles in important biological processes like cell proliferation, differentiation, apoptosis, and tumorigenesis through post-transcriptional regulatory networks. Traditional detection technologies (such as quantitative Polymerase Chain Reaction (qPCR), Northern blot, and microarrays) exhibit high sensitivity, yet they rely on expensive thermocycling equipment, involve cumbersome procedures, and are time-consuming, making it difficult to meet the demands of clinical point-of-care testing. In recent years, isothermal amplification technologies, represented by loop-mediated isothermal amplification (LAMP), Rolling Circle Amplification (RCA), Strand Displacement Amplification (SDA), and Exponential Amplification Reaction (EXPAR), have emerged as a research frontier in the field of miRNA detection, by virtue of their advantages including elimination of thermal cycling requirements, fast reaction kinetics, high sensitivity, and simple operation. This study aims to systematically organize the core principles, latest advancements, and integrated innovation patterns of isothermal amplification-based miRNA detection technologies, and provide support for the development of miRNA detection platforms suitable for clinical point-of-care testing.

1. INTRODUCTION

A class of endogenous non-coding single-stranded RNAs with a length of roughly 18–25 nucleotides (nt) is known as microRNAs (miRNAs). They participate in important biological processes like cell proliferation, differentiation, apoptosis, and metabolic reprogramming by precisely regulating gene expression networks at the post-transcriptional level through sequence-specific complementary binding to the 3' Untranslated Regions (3'-UTRs) of target mRNAs [1-4]. Numerous studies conducted in recent years have shown that dysregulation of miRNA expression profiles is closely linked to the development, progression, and drug resistance of a number of human diseases, especially malignant tumors. As a result, miRNAs have emerged as extremely promising liquid biopsy biomarkers and precision medicine targets [3, 5]. Nonetheless, miRNAs have intrinsic properties such as a short sequence, low abundance, high sequence homology within the family, and vulnerability to RNase degradation in physiological fluids. Achieving highly sensitive, highly specific, quick, and convenient detection is extremely difficult due to these characteristics [6, 7].

High-throughput sequencing (NGS), microarray chips, Northern blotting, in situ Hybridization (ISH), and qPCR are currently the most widely used miRNA detection techniques [8, 9]. While NGS enables profiling of the entire transcriptome's miRNA repertoire in a single run, it has drawbacks such as expensive instruments, complex data analysis, and sensitivity limitations that depend on sequencing depth [10]. Despite their advantage of parallel detection, microarray chips suffer from a narrow dynamic range, significant batch-to-batch variations, and challenges in achieving absolute quantification [11, 12]. Even though northern blotting is regarded as one of the “gold standards” for its direct ability to separate mature miRNAs from their precursors, its application in high-throughput settings is limited due to laborious procedures, days-long turnaround time, and detection limits typically in the picomolar range [13, 14]. While probe permeability issues and background signal interference restrict the quantitative accuracy of ISH, this technique can visualize the spatial distribution of miRNAs within tissues [15]. As for qPCR, it has been widely used due to its high sensitivity; however, given the short sequence of miRNAs, it must rely on stem-loop or tailing reverse transcription strategies, which involve cumbersome procedures, require thermocyclers, and are prone to introducing biases. Additionally, in low-abundance samples, issues such as low amplification efficiency and insufficient specificity tend to arise [16-18]. Therefore, to address these current limitations, efforts should be made to develop new technological platforms capable of performing amplification and detection under isothermal conditions while eliminating the need for thermocyclers.

Due to its benefits, which include mild reaction conditions, low equipment requirements, short reaction times, high sensitivity, and strong specificity, Isothermal Amplification (IA) technology has become a popular alternative in the field of miRNA detection in recent years [19]. For the quick and accurate measurement of miRNAs, methods such as Rolling Circle Amplification (RCA), Strand Displacement Amplification (SDA), Exponential Amplification Reaction (EXPAR), Loop-Mediated Isothermal Amplification (LAMP), and enzyme-free catalytic amplification systems have been reported one after the other [20-22]. Nevertheless, non-specific amplification, primer-dimer interference, and complicated probe design are still issues with single isothermal amplification techniques. In order to overcome these obstacles, scientists have combined isothermal amplification with CRISPR/Cas systems, functionalized nanoprobes, and multi-signal amplification units to create portable, all-in-one, ultra-high-sensitivity biosensing platforms. This integration offers new methods for implementing miRNA in precision medicine and Point-of-Care Testing (POCT) [23, 24]. Against this backdrop, this study focuses on organizing the evolution of mainstream technical principles, key performance breakthroughs, and multi-technology integration mechanisms of isothermal amplification technology in miRNA detection, aiming to provide targeted theoretical references and practical guidance for optimizing the highly sensitive, specific, portable, and integrated miRNA detection technical schemes that meet the needs of point-of-care testing, as well as their clinical translation pathways.

2. TECHNIQUES FOR ISOTHERMAL AMPLIFICATION-BASED MIRNA DETECTION

2.1. Rolling Circle Amplification(RCA)

Due to its high sensitivity, high specificity, and “one-pot” reaction feature that does not require thermal cycling, Rolling Circle Amplification (RCA) technology has emerged as one of the most researched isothermal amplification techniques for miRNA detection [25]. Its fundamental idea entails two crucial steps, as shown in Fig. (1): To create a closed circular template, T4 DNA ligase first circularizes the linear padlock probe with the help of miRNA. Under isothermal conditions, Phi29 DNA polymerase then carries out rolling circle extension by utilizing its strand displacement activity and processivity. This achieves exponential signal amplification by producing long single-stranded DNA (ssDNA) products made up of hundreds to thousands of repeating units [26-28].

Fig. (1).

Schematic diagram illustrating the principle of RCA for miRNA detection.

For endpoint detection, traditional RCA uses gel electrophoresis, which has a low throughput and low sensitivity. By combining various signal transduction platforms, such as fluorescence, colorimetry, electrochemistry, and functionalized nanomaterials, researchers have improved the detection performance and clinical applicability of RCA technology in recent years [29, 30]. Song et al. developed a method called Palindromic Hyperbranched Rolling Circle Amplification (PH-RCA), in which a dumbbell probe changes shape when miRNA is present to start RCA. Palindromic sequences then self-assemble to create a multibranched structure. This allows for accurate discrimination of single-base mismatches and attains a detection limit of 18.5 aM [31]. Self-Primed Extension (SPE) and RCA were combined by Shao et al. [32]. A strand-displacing polymerase continuously extends the signal after biotin-labeled RCA products are captured on a streptavidin-coated microplate surface. With this method, the detection limit is as low as 10 fM, and it shows good tolerance to the serum matrix. CRISPR/Cas signal amplification technology has been used by researchers to further increase the sensitivity of miRNA detection. Zhou et al. [33], for example, created an integrated system called T-ERCA that combines a CRISPR/Cas12a trans-cleavage system with dumbbell probe-triggered exponential RCA: By acting as an activator, the lengthy ssDNA product produced by RCA causes Cas12a to cleave fluorescent reporter probes in large quantities, facilitating secondary signal amplification. As shown in (Fig. 2), this successfully profiles miRNA expression in various cell lines by lowering the detection limit to 0.31 fM.

Fig. (2).

(A, B). T-ERCA combined with CRISPR/Cas12 system: detection schematic.

Despite RCA's encouraging potential in miRNA detection, non-specific amplification and false positives caused by extremely similar sequences continue to be major obstacles to its clinical application [34, 35]. Research shows that lock probes can mis-circularize and produce increased background signals when sequence differences between members of the miRNA family are only 1–2 nucleotides [36]. In the future, a “Padlock Design-and-Operation Checklist” will be implemented to eliminate mis-circularization caused by homologous sequences: a 2-nt gap is positioned in the variable region and ligated with SplintR at 37R at 37on caused by homologous sequences: a 2-nt gapdT or C3-spacer is introduced to block off-target extension, reducing the background by≥5-fold; three single-base-mismatch synthetic targets must exhibit ligation efficiencies ≤ 5% (ΔCt ≥ 4.3), with PLC/NLC data disclosed in the main text; optional CRISPR/Cas13 secondary cleavage can further drive the false-positive rate to ≤ 1%. Coupling these measures with a microfluidic-RCA chip will provide an error-controlled and fully reproducible all-in-one solution for POCT and precision medicine.

Future developments could further lower false-positive rates and move RCA technology closer to Point-of-Care Testing (POCT) and precision medicine applications by improving the secondary structure of lock probes, introducing programmable nucleases (like CRISPR/Cas13) for real-time correction of non-specific products, and creating integrated microfluidic-RCA chips.

2.2. Loop-mediated Isothermal Amplification (LAMP)

First described by Notomi et al. [37]. in 2000, loop-mediated isothermal amplification (LAMP) is a “one-pot” method that can exponentially amplify nucleic acids at a fixed temperature (55–65°C). Targets are amplified by 109–1010 times in 15–60 minutes using Bst DNA polymerase with strand-displacing activity, which is characterized by the use of 4–6 specific primers (F3/B3, FIP/BIP, and optional loop primers LF/LB) that recognize 6–8 distinct regions of the target sequence [38-40]. Its canonical reaction pathway is as follows, as shown in (Fig. 3): FIP anneals to the template and extends to form a stem-loop structure; the F3 primer then initiates strand displacement; multiple cycling produces dumbbell-shaped intermediates with terminal stem-loops; and finally, a large number of “cauliflower-like” long-stranded DNA products with repeating target sequences are produced.

Fig. (3).

Schematic illustration of the LAMP working principle.

However, because miRNA is only about 22 nt long, it is challenging to directly meet the 100–200 bp targets that LAMP typically requires [41]. Two approaches have been suggested by researchers to deal with this: (1) Stem-loop primer adaptation: Park Jee et al. [42] created chimeric primers that allow miRNA to directly initiate LAMP as a “trigger strand” without reverse transcription. The 5' end of these primers is complementary to miRNA, while the 3' end contains LAMP inner primer sequences. Similarly, the target miRNA was used by Ye et al. [43] to self-fold and extend with DNA in order to prime LAMP reactions. (2) Dumbbell template bridging: In a miRNA-LAMP system (Fig. 4), SplintR ligase joins miRNA-mediated Linker A/B probes into dumbbell-shaped templates, which are then amplified by Bst 2.0 polymerase. This technique was developed by Wu et al. [44]. This achieves seven-order linear quantification within 100 fM–10 nM, with a detection limit as low as 100 fM, when used in conjunction with EvaGreen real-time fluorescence monitoring.

Fig. (4).

Schematic illustration of miRNA detection using the miRNA-LAMP system. Reproduced from Wu et al., Int. J. Anal. Chem., 2023, under the terms of the Creative Commons Attribution 4.0 International License (CC BY 4.0) [44].

LAMP is being combined with several cutting-edge technologies to create multimodal cooperative amplification strategies in an effort to further reduce non-specific amplification and improve sensitivity [45, 46]: (1) RCA-LAMP cascade: Cao et al. [47] achieved a detection limit of 500 aM by using RCA-generated long single-stranded products as LAMP templates. The complexity of the LAMP primer is greatly decreased by this RCA pre-amplification.(2) LAMP-CRISPR/Cas12a cascade: Following LAMP amplification, Mao's group [48] added Cas12a using fluorescent reporter probes. By using Cas12a's trans-cleavage activity to amplify secondary signals, they were able to achieve single-base resolution and a detection limit of 0.1 fM.

In conclusion, even though LAMP offers a lot of benefits for point-of-care miRNA detection, the following issues still need to be fixed: (1) Non-specific amplification and primer dimerization: Erroneous amplification is easily triggered by minor sequence variations within miRNA homologous families. (2) Multiplex detection limitations: The implementation of multiplex LAMP is limited by crosstalk between complex primer sets and multicolor fluorescent probes. (3) Bottlenecks in field deployment: Because conventional LAMP still uses benchtop isothermal instruments, it must be further integrated with portable fluorescence readers, paper-based devices, or microfluidic chips. Future studies should concentrate on the following areas: (1) high-specificity primer design guided by machine learning, (2) improved mismatch discrimination via artificial base pairs or locked nucleic acid (LNA) modifications, (3) Synchronous multi-miRNA detection made possible by dual-compartment microfluidic-LAMP chips, and (4) Low-cost fluorescence/colorimetric readout systems that work with smartphones to enhance LAMP clinical translation in environments with limited resources.

2.3. Strand Displacement Amplification (SDA)

Strand-displacing DNA polymerases and restriction endonucleases (also known as nicking enzymes) work in concert to achieve exponential target amplification in Strand Displacement Amplification (SDA), an isothermal nucleic acid amplification technique that can be used at 55–65°C [49, 50]. Target miRNA anneals to a hairpin template, the enzyme nicks specific sites, polymerase facilitates extension and downstream strand displacement, and the single-stranded DNA template regenerates cyclically. This is its basic workflow. In just a few minutes, this cascade produces a large number of double-stranded products [51, 52]. Currently, the main miRNA detection strategies based on SDA include (Note: The limit of detection (LOD) data, disease discrimination cases, and clinical application explorations of the following strategies are preliminary research results):(1) A “one-pot” exponential SDA (E-SDA) system was created by Shi et al. [53]. in which two rounds of circular strand-displacement cycles (polymerization-cleavage-displacement, Fig. 5) are initiated by target miRNA. It greatly outperforms traditional linear SDA, achieving signal amplification in 90 minutes with a detection limit of 16 zmol without the need for reverse transcription. (2) CRISPR-coupled SDA: Using miRNA as an SDA initiator, Chi et al. [54] developed a CRISPR-Cas14a-coordinated approach (Fig. 6). G-rich single-stranded DNA is produced by target miRNA driving SDA, which triggers Cas14a's trans-cleavage activity for secondary signal amplification. In 30 μL serum samples, this technique successfully distinguishes cholangiocarcinoma patients from healthy controls, achieves an ultralow detection limit of 0.68 aM, and accurately separates miRNA-21 from single-base mutants, exhibiting remarkable clinical utility. (3) Micro-nano integrated SDA: Researchers have combined SDA with RCA cascades, magnetic nanoparticles (MNPs), or microchips to further increase sensitivity and throughput. For example, Xie et al. [55] enabled high-throughput parallel analysis by combining microchip electrophoresis with SDA and achieving a detection limit of 0.02 fM for miRNA-21 in a 5 μL reaction volume. The MNPs-SDA electrochemical sensor was created by Xu et al. [56] and successfully applied to the sera of breast cancer patients. MNP-conjugated capture probes enrich target miRNA in complex samples, achieving a sensitivity of 0.5 fmol·L⁻¹. Zhou and colleagues [57] designed an RCA-SDA cascade: By producing long single-stranded templates for exponential SDA amplification, RCA pre-amplification enables femtomolar-level multiplex detection of miRNAs linked to colorectal cancer with noticeably higher specificity. Additionally, creating target-specific primers and probes is essential to raising the precision of miRNA detection. High-fidelity primers and probes significantly improve target miRNA selective recognition in SDA systems while lowering non-specific amplification [58, 59]. According to Li's research, better primer design increases specific binding to target miRNAs, leading to increased sensitivity and specificity [60].

Fig. (5).

Schematic illustration of miRNA-triggered Exponential SDA (E-SDA).

Fig. (6).

Schematic illustration of miRNA detection mediated by the Cas14-SDA system.

The following bottlenecks still exist despite SDA's benefits, which include minimal reaction systems and thermal cycling independence:(1) strand-displacement byproduct accumulation that increases background signals; (2) reduced nicking enzyme activity when targeting complex secondary structures or miRNAs with high GC content; and (3) instability in detection as a result of serum RNase-mediated miRNA degradation. Future developments could concentrate on: (1) improving mismatch discrimination and probe stability through Locked Nucleic Acid (LNA) or 2'-O-methyl modifications, (2) creating integrated microfluidic-SDA chips that would allow for completely automated sample-to-answer workflows, and (3) constructing Cas14a-SDA dual-target logic gates to enable multiplex miRNAs to be precisely discriminated at the same time. With these tactics, SDA is well-positioned to advance from bench to bedside as a reliable tool for miRNA detection in liquid biopsy for cancer.

2.4. Enzyme-free Amplification Technologies

Enzyme dependency is totally eliminated by enzyme-free amplification technologies, which use entropy-driven strand displacement for signal amplification and programmable self-assembly of DNA molecules. These methods are important innovations for ultrasensitive miRNA detection because they combine benefits like low cost, high stability, and simplicity of automation [61-65]. Catalytic Hairpin Assembly (CHA), Hybridization Chain Reaction (HCR), and Entropy-Driven Amplification (EDA) are examples of contemporary mainstream techniques. By suppressing background noise and greatly expanding the linear detection range, these methods can be used singly or in combination.

2.4.1. Catalytic Hairpin Assembly (CHA)

The target miRNA is used in Catalytic Hairpin Assembly (CHA) to cause two metastable hairpin probes to cascade hybridize. To achieve “target-catalyzed” signal amplification, a fresh round of conformational switching is catalyzed by each target binding event [66, 67]. Dong et al. [68] used flow cytometry to detect Fluorescence Resonance Energy Transfer (FRET) signals after anchoring CHA reactions on magnetic bead surfaces. The detection of miRNA-155 at 2 pM was made possible by the significant reduction of fluorescence background caused by the micrometer-scale spatial confinement of beads (Fig. 7). Research shows that CHA can detect miRNA at picomolar concentrations with high sensitivity [69], which is frequently impossible with traditional techniques. To further improve detection sensitivity and specificity, CHA can be combined with fluorescence imaging or electrochemical sensors [70].

Fig. (7).

Schematic illustration of the miRNA detection strategy using CHA-coupled flow cytometry (FRET).

2.4.2. Hybridization Chain Reaction (HCR)

With the help of miRNA-induced alternating polymerization of two DNA hairpins (H1/H2), Hybridization Chain Reaction (HCR) produces double-stranded DNA nanowires that are n times longer and accomplish geometric signal amplification. Researchers have investigated new approaches such as localized HCR, migrating HCR, and multibranched HCR to improve HCR efficiency [71-73], which has greatly increased detection throughput and sensitivity [74].

Among the notable implementations are the HCR-driven fiber-embedded microfluidic chip (HCR-FMB) built by Lu et al. [75]. It detects serum let-7a in 6 minutes with a detection limit of 100 pM and remarkable specificity by combining FRET and evanescent wave fluorescence. Tetrahedral DNA framework (TDN)-assisted CHA-HCR dual amplification was used by Zhang et al. [76] to create a cascade amplification system (Fig. 8). When used for in situ imaging of let-7a in living cells, the TDN nanostructure spatially colocalizes CHA and HCR reactions, shortening reaction pathways to reach an ultralow detection limit of 8.51 pM.

Fig. (8).

(a-c). Schematic illustration of the working principle for detecting let-7a using the TDN-CHA-HCR system.

2.4.3. Entropy-Driven Amplification (EDA)

Entropy-Driven Amplification (EDA) uses strand displacement reactions powered by “entropy gain and free energy reduction” to detect miRNAs in an enzyme-free, highly specific manner [77, 78]. The following are important EDA-based miRNA detection techniques: (1) Surface-enhanced EDA: Cai et al. [79] designed two EDA reaction layers on the surface of gold nanoparticles. When combined with fluorescence quenching/recovery mechanisms, this allows for intracellular imaging of miRNA-21 that can differentiate between healthy and cancerous cells; and (2) CHA-EDA cascade: The entropy-driven reaction was started by Huang et al. [80] using CHA products as fuel strands (Fig. 9). The sensitivity of miR-208a detection is increased to 1.3 fM by this two-stage amplification in a single tube, which is a 100-fold improvement over traditional CHA.

Fig. (9).

Schematic illustration of the working principle for the enzyme-free cascade amplification biosensor detecting microRNA using the EDA-CHA system.

Despite the impressive performance of enzyme-free amplification technologies, they still face bottlenecks such as increased background signals caused by the adsorption of hairpin probes by non-target nucleic acids in complex samples, reduced single-base discrimination due to hairpin cross-reactivity, and highly temperature- and salt concentration-dependent kinetics of EDA reactions, resulting in suboptimal field stability that requires improvement. Future strategies may involve introducing locked nucleic acid (LNA) or peptide nucleic acid (PNA) modifications to enhance probe specificity, developing AI-assisted hairpin design platforms, and integrating with microfluidic platforms. These approaches will help overcome the barriers in complex sample analysis and clinical scenarios, accelerating the translation of miRNA biomarkers into point-of-care diagnostics and precision medicine.

2.5. Exponential Amplification Reaction (EXPAR)

Short DNA templates, nicking enzymes, and strand-displacing DNA polymerases are used in the Exponential Amplification Reaction (EXPAR) to accomplish exponential amplification at isothermal temperatures (55–60°C) [81, 82]. Its template has an Nt—BstNBI nicking site along with two repeating sequences. A complementary strand is produced by polymerase extension when miRNA hybridizes with the template's 3' end. Following this strand's cleavage by the nicking enzyme, new primers are created, starting additional extension and downstream strand displacement. Target miRNA is amplified exponentially by this “extension-cleavage-displacement” cycle (Fig. 10) [83]. EXPAR has become a potent tool for point-of-care miRNA detection due to its reverse transcription-free operation, ultrahigh sensitivity (zmol level), and quick kinetics (<30 min).

Fig. (10).

(a and b). Schematic illustration of microRNA detection using EXPAR.

EXPAR performs exceptionally well in miRNA detection, especially in terms of reaction efficiency and sensitivity [84-86]. Studies have demonstrated that portable detection platforms integrated with EXPAR exhibit application potential in early-stage feasibility studies. They enable the detection of multiple miRNAs in urine samples within 30 minutes, with an accuracy of 83% [87]. Pan et al. [88] used thermodynamic template optimization to achieve a 90% reduction in background amplification and a detection limit of 10 zmol for let-7a. By using hairpin templates to block non-specific interactions, Peng et al. [89] created a self-priming EXPAR (SPEXPAR)-lateral flow colorimetric platform that allows for quantitative miRNA-155 detection through gold nanoparticle (AuNP) color transitions in 35 minutes over a linear range of 0.41 pM–100 nM.

Despite the many advantages of EXPAR technology, background amplification signals remain a major challenge in EXPAR reactions [90]. The main sources of non-specific background amplification signals are: (1) primer-mediated non-specific amplification, (2) template-induced non-specific amplification, and (3) DNA polymerase-mediated non-specific amplification. The main causes of EXPAR background signals and improved measures to suppress background amplification signals will be discussed. The high non-specific background of EXPAR stems from primer mismatches, template self-hybridization, and non-specific extension by polymerases. The following text reviews non-specific suppression strategies from three aspects: molecular design, chemical modification, and system engineering.

2.5.1. Primer Optimization

Due to their short sequence length and high GC content, miRNAs are more likely to exhibit partial complementarity with non-target sequences, which can result in mismatch extension and negatively impact the sensitivity and reliability of detection [91]. Moreover, non-specific binding greatly increases background signals when primers have sequence similarities with off-target regions [92]. Non-specific binding can be significantly decreased by using machine learning for secondary structure prediction, optimizing the design of miRNA primers (GC content, length, and target match specificity), or adding LNA/2′-OMe modifications to improve ΔTm.

2.5.2. Template Engineering

One of the main sources of EXPAR background signals is non-specific template amplification, which results from transitory hybridization at template 3'-ends [93, 94]. When the 3'-termini of two templates come into contact during nucleic acid amplification, complementary base pairing may result in the formation of unstable duplex structures. Furthermore, templates have the ability to self-fold into hairpin patterns, which results in the temporary production of duplexes at their 3'-termini. By activating DNA polymerase, these structures behave as pseudo-priming sites, starting amplification without the need for valid initiator strands. Fig. (11) [95] shows the four main unintentional amplification paths of autonomous templates. Importantly, polymerase extension can be primed by complementary sequences at 3'-ends without the need for particular primers, increasing background noise and lowering detection accuracy [96].

Fig. (11).

(A) Inter-template 3'-end hybridization. (B) Intra-template 3'-terminal hairpin formation. (C) Transient template-template interactions. (D) Hybridization between template fragments and intact templates.

In order to decrease EXPAR background signals, structural modification techniques that prevent the non-specific amplification of EXPAR templates primarily consist of: (1) Template end-capping technology: this technique stops non-specific amplification by altering the nucleic acid template's 3' end to stop polymerase extension. Research has demonstrated that closing the 3' end with inverted T, C3 Spacer, PNA, and other modifications can not only successfully stop polymerase extension but also improve template stability, increasing the detection's sensitivity and specificity [82, 97, 98]. Furthermore, a study conducted by Chen et al. [99] showed that modifying the template's 5' end with biotin can lower its melting temperature (Tm), which enables miRNA to preferentially bind to the 3' end and extend, minimizing template loss and increasing detection efficiency. (2) Secondary Structure Design: Nonspecific signals can be efficiently suppressed by the secondary structure design of the template, particularly the hairpin structure. Using the hairpin conformation to prevent template self-extension, Guo et al. [100] created an exp-Hairpin template with embedded cut sites in the stem region. This resulted in a theoretical amplification efficiency of >2^N and a detection limit of miRNA-21 at 5.84 fM, greatly increasing the reaction's specificity (Fig. 12). (3) Chemical modification techniques: Thioester bonds, methylation, phosphorylation, and glutaraldehyde treatment can all improve the sensitivity and accuracy of detection by increasing template rigidity and decreasing non-specific binding [101-103].

Fig. (12).

Schematic illustration of the working principle for miRNA detection using the exp-Hairpin strategy.

2.5.3. System Engineering

In order to limit the template's nonspecific amplification, the system engineering approach primarily entails adding additional substances to the reaction system. Primarily: (1) Additive strategy: single-stranded binding proteins (SSB) can reduce the background by three orders of magnitude; 5–10% glycerol or DMSO can reduce π-π stacking between templates [95, 104]. (2) Nanomaterial shielding: By using π-π interactions to adsorb free templates, Graphene Oxide (GO) stops polymerase from binding non-specifically, enabling detection at the 5 aM level with a sensitivity that is 10⁴ times higher than that of conventional EXPAR [105]. (3) 1,6-hexanediol and two-dimensional CoOOH nanosheets can prevent template self-assembly, which reduces background signal by more than 95% [106].

2.5.4. DNA Polymerase-mediated Non-specific Amplification

Random primer formation and extension are crucial steps in DNA polymerase's de novo synthesis capability, which is based on structural characteristics and catalytic mechanisms and is the main source of non-specific amplification [107, 108]. The ability of DNA polymerases to displace existing DNA strands during nascent strand synthesis, known as strand-displacing activity, is the primary driving force behind the synthesis of random primers in the absence of target sequences.

As an illustration, the enzyme Bst Exo-polymerase has a high affinity for binding purine-rich DNA motifs. In particular, by promoting polymerase-primer interactions, dG at the 3'-terminus of primers encourages random primer formation [109]. Random primers are extended to create new DNA strands after polymerase binding. Any section of the template DNA could act as a primer extension initiation site during this procedure, resulting in non-specific amplification [110, 111]. We advise using modified nucleotides with greater fidelity and optimizing primer design to steer clear of dA/dG-rich 3'-termini in order to reduce random primer generation. These methods greatly improve the reliability and specificity of amplification.

In summary, the EXPAR technology shows great potential in the field of nucleic acid detection, but its non-specific amplification remains a critical bottleneck to be addressed. Based on the aforementioned strategies, a preliminary background suppression protocol can be established: using 3'-end blocked hairpin templates (containing a nicking site) in combination with Nt.BstNBI nuclease and Bst Exo⁻ polymerase, with the reaction system supplemented with 5% glycerol and single-stranded binding protein (SSB) to reduce non-specific interactions. It specifies that the fluorescence rise time of the No-Template Control (NTC) should be ≥60 minutes (delayed by ≥30 minutes compared to the target), with a background slope ≤0.02 RFU/min (within 20-60 minutes). For crude matrices such as serum and urine, blinded validation is required, specifying processing details (e.g., serum inactivation at 56°C + centrifugation), blinded detection accuracy (≥85% for serum, ≥80% for urine), and concentrations and impacts of interferents (e.g., hemoglobin), thus providing replicable standards for the stable application of EXPAR.

Future development of EXPAR will focus on: (1) AI-based template-primer co-design platforms to achieve single-base resolution, (2) integration of microfluidics with EXPAR chips to enable sample-in-result-out POCT modes, and (3) multiplex target EXPAR encoding systems to meet clinical needs for simultaneous quantification of multiple miRNAs. With the maturation of background suppression strategies and device miniaturization, EXPAR is expected to achieve widespread translation in the fields of early tumor screening, prognostic monitoring, and home self-testing.

3. DISCUSSION

3.1. Core Value of Isothermal Amplification in miRNA Detection

Isothermal amplification technologies, with core advantages of “no need for thermal cycling, rapid reaction, and high sensitivity,” have significantly broken through the technical bottlenecks of conventional qPCR in miRNA detection, providing a new methodological basis for early cancer screening, companion diagnosis, and prognostic monitoring. This review systematically summarizes the principal innovations, performance limits, and clinical validation of strategies such as RCA, LAMP, SDA, EXPAR, and enzyme-free amplification (CHA/HCR/EDA) in miRNA quantification (Table 1). Their common advantages are reflected in: (1) Simple operation, achieving detection limits at the aM–zmol level with a single-temperature reaction, (2) Seamless coupling with multidisciplinary technologies (e.g., CRISPR, nanomaterials, microfluidics) to realize secondary signal amplification and on-site detection, and (3) Good compatibility with short, low-abundance, and highly homologous miRNAs.

Table 1.

Characteristics of different isothermal amplification methods.

Method	Type	Target	Reaction Volume	Detection Time	Detection Range	LOD	References
RCA	CRISPR-Cas12a	miR-21	110 μL	70 min	100 pM-10 nM	8.1 pM	[27]
	DNA(HD)-RCA	miR-21	30 μL	90 min	1 aM-1 pM	39.7 aM	[28]
	PH-RCA	miR-141	N/A	30 min	100 aM-1 nM	18.5 aM	[31]
	T-ERCA	miR-155	20 μL	100 min	1 fM-5 nM	0.31 fM	[33]
LAMP	Ligation-LAMP	miR-155	50 μL	110 min	10 fM-100 pM	15 fM	[39]
	RT-THSP-LAMP	miR-152	20 μL	120 min	10 fM-3 nM	10 fM	[43]
	miRNA-LAMP	miR-200a	30 μL	180 min	100 fM-5 nM	100 fM	[44]
	CAL-LAMP	let-7a	200 μL	73 min	100 aM-100 fM	0.1 fM	[48]
SDA	B-HCR	miR-21	20 μL	150 min	0-60 nM	11.3 pM	[49]
	SDA	let-7a	10 μL	90 min	0.1 zmol-1.0 pmol	16 zmol	[53]
	Cas14SDA	miR-21	30 μL	60 min	0.68-50 pM	680 fM	[54]
	CRISPR/Cas12a	miR-21	20 μL	150 min	0.5 fM-100 pM	0.5 fM	[56]
CHA	split-G4-CHA	miR-21	20 μL	70 min	10 fM-100 nM	37 fM	[66]
	FCM-CHA	let-7a	10 μL	180 min	0.7 pM-300 pM	0.7 pM	[68]
	FA-CHA	miR-21	N/A	240 min	1.0-19.0 nM	52.0 pM	[67]
	ECL-CHA	miR-21	5 μL	120 min	1 aM-1 nM	0.02 aM	[70]
HCR	CRISPR/Cas12a	miR-21	N/A	60 min	250 fM-10,000 pM	17.6 amol	[72]
	HCR-FMB	let-7a	20 μL	6 min	3.0-1000 nM	100 pM	[75]
	TDN-CHA-HCR	let-7a	N/A	60 min	0.05-5 nM	8.51 pM	[76]
EDA	CRISPR/Cas13a	miR-17	20 μL	45 min	1 fM-100 pM	1.05 fM	[77]
	AuNPs-EDA	miR-21	100 μL	120 min	5 pM-25 nM	5 pM	[79]
	CHA-EDA	miR-21	N/A	150 min	50 fM-1 pM	1.3 fM	[80]
EXPAR	EXPAR	let-7a	10 μL	30 min	0.1 zmol-1.0 pmol	0.1 zmol	[83]
	SPEXPAR	miR-155	10 μL	40 min	10 fM-1 μM	0.41 pM	[89]
	exp-Hairpin EXPAR	let-7a	30 μL	34 min	10 fM-10 nM	5.84 fM	[100]
	GO/SSB	trigger	20 μL	60 min	5 aM-50 fM	5 aM	[105]
	CoOOH	miR-122	20 μL	N/A	50 aM-500 fM	151 ymol	[106]

3.2. Limitations of Current Technologies

Despite the significant advantages of isothermal amplification technologies in miRNA detection, several challenges remain: (1) Non-specific amplification and cross-reactivity have not been fully eradicated, resulting in prominent background noise in the low-copy number range, (2) In multiplex parallel detection of targets, complex interactions between primers and templates reduce quantitative accuracy, (3) RNases, inhibitors, and pH fluctuations in complex biological matrices (e.g., whole blood, tissue lysates) easily interfere with amplification efficiency, and clinical reproducibility requires further verification, and (4) Most existing platforms remain in the stage of laboratory proof-of-concept, lacking complete solutions for deep integration with commercial POCT devices.

Regarding the core issues directly related to the standardization of sample processing and data reliability among the aforementioned challenges (such as interference from complex matrices, insufficient reproducibility caused by variations in sample pretreatment, and inconsistent quantitative calibration methods), a series of Standardized Operating Procedures (SOPs) adapted to isothermal amplification technologies has been initially developed based on existing studies. These SOPs provide practical and implementable solutions for enhancing the robustness of detection. Tables (2 and 3) below systematically sort out key operational points and technical adaptability from the perspectives of pre-analytical variable control and data standardization, respectively, offering references for experimental design and quality control in clinical translation.

Table 2.

Operating specifications for pre-analytical variable control.

Module	Key Operations	Critical Control Points	Compatible Isothermal Technologies	Limitations	Application Scenarios	References
Pre-analytical Variable Control	1. Sample collection and anticoagulation	- Blood/serum: EDTA tubes; centrifuge 3000×g, 4°C, 10 min (≤30 min post-collection). Avoid heparin (47% enzyme inhibition). - Urine/CSF: RNase-free tubes + 0.1% DEPC.	- LAMP (heparin-tolerant); - RCA (strict anticoagulant control).	- EDTA reduces miR-451a recovery (~8%); - Heparin impairs SDA (23% efficiency loss).	- Clinical blood/fluid testing; - Multicenter studies.	[112, 113]
	2. Storage and freeze-thaw control	- Isolate via ultracentrifugation (100,000×g, 4°C, 70 min) or ExoQuick. - Lyse with 0.5% Triton X-100 (release rate: 62%→91%).	- EXPAR (degradation-sensitive); - LAMP (mild tolerance).	- miR-16 declines 12–15% after 6-month -80°C storage ; - Prolonged ice thawing ↑RNase (2.3×).	- Biobank samples; - Low-abundance miRNA detection.	[114]
	3. miRNA extraction and lysis	- Isolate via ultracentrifugation (100,000×g, 4°C, 70 min) or ExoQuick. - Lyse with 0.5% Triton X-100 (release rate: 62%→91%).	- RCA (requires efficient lysis); - SDA (reliant on free miRNAs).	- Ultracentrifugation loses ~30% exosomes (vs. 15% with kits); - Triton >0.5% ↓LAMP (18%).	- Liquid biopsies (tumor markers); - Extracellular vesicle research.	[115]
	4. Matrix compatibility and inhibition checks	- Complex matrices: 1:10 dilution or inhibitor columns. - Simple matrices: 1× TE (pH 8.0) + 0.1% SDS. - Inhibitor test: ΔCt <2 for miR-16.	- LAMP (anti-interference); - EXPAR (requires ≥1:50 dilution).	- Complex matrices (whole blood, feces); - Novel matrices (saliva).	- Complex matrices (whole blood, feces); - Novel matrices (saliva).	[116]

Table 3.

Operating specifications for standardization strategies.

Module	Key Operations	Critical Control Points	Compatible Isothermal Technologies	Limitations	Application Scenarios	References
Normalization Strategies	1. Exogenous spike-in calibration	Use non-human miRNAs (e.g., cel-miR-39) at 50 fM (added post-collection). Require ≥70% recovery.	- All (match target efficiency); - Prefer LAMP/SDA (broad linear range).	- Multicenter studies; - Assay validation.	- Multicenter studies; - Assay validation.	[117]
	2. Reference miRNA selection	Candidates: miR-16/miR-451a (serum); miR-122/miR-93 (tissues). Select via geNorm (M-value <0.5).	Single-target detection; avoid miRNA families (cross-amplification).	- Clinical typing; - Tissue expression analysis.	- Clinical typing; - Tissue expression analysis.	[118]
	3. Global mean normalization (multiplex)	For ≥5 miRNAs: calculate mean Ct; exclude outliers (±3 SD). Normalization: target Ct – mean Ct.	Multiplex digital LAMP; chip-based amplification (≥10 miRNAs).	- High-throughput screening; - Novel samples.	- High-throughput screening; - Novel samples.	[119]
	4. System quality control	RNase control: DEPC consumables + 0.1 U/μL RNasin. - NTC signal <10% of target; CV <15% (3 replicates).	All (EXPAR/RCA: contamination-sensitive); automated platforms.	- Clinical lab QC; - Automated POCT.	- Clinical lab QC; - Automated POCT.	[120, 121]

3.3. Future Perspectives

In the future, leapfrog improvements in isothermal amplification miRNA detection can be made from the following dimensions: (1) Molecular design driven by algorithms: A “sequence-structure-performance” prediction model is developed using AI-assisted thermodynamic calculations and machine learning to achieve the automatic and ideal design of primers, templates, and probes, greatly reducing the experimental iteration cycle, (2) The multiplexing strategy at the system level: The requirements of tumor molecular subtyping and drug resistance monitoring are met by combining isothermal amplification with a high-throughput decoding platform to achieve absolute quantification of >5 miRNAs in a single reaction using programmable barcodes, microarray partitioning, or digital microdroplet technology. For example, an isothermal LAMP system employing 55-µm droplets (λ = 0.06) was used to compartmentalize individual templates; the predicted droplet collision rate was ≤ 0.55%. A 12-dimensional code space generated by four orthogonal fluorophores × three intensity levels was narrowed to six dimensions that cover liver-cancer-related miRNAs. After AI-based primer filtering, the decoding error remained < 1% and serum validation yielded r = 0.93 versus qPCR, establishing a reproducible, >5-plex, isothermal multiplexing paradigm, (3) Anti-interference system of clinical quality: To guarantee the robustness of detection in complex samples, an integrated sample pre-processing module (rapid lysis, inhibitor adsorption, and RNA enrichment) was created. Additionally, internal reference correction and real-time quality control probes were added, and (4) End-to-end automated platform: It deeply integrates isothermal amplification (with precise temperature control and minimal fluctuations), smartphone-based optical reading (equipped with a custom dark box and specialized optical filter, achieving a significantly enhanced optical signal-to-noise ratio), cloud-based AI interpretation (with high accuracy), and blockchain-based data management (with an extremely low risk of tampering). Through fluorescent microsphere calibration targets and an auto-exposure locking mechanism, it effectively controls device variability, thereby constructing a full-link closed-loop system encompassing “sample-result-report.” This platform features a short total processing time and greatly simplified operational procedures, and is truly applicable in primary hospitals, emergency vehicles, and home self-testing scenarios.

CONCLUSION

In conclusion, the transition from “methodological innovation” to “clinical transformation” in isothermal amplification technology has reached a critical window. In the future, it is anticipated that collaborative innovation in algorithms, materials, chips, and data science will advance miRNA detection from the lab to the final mile of public health management.

LIST OF ABBREVIATIONS

EXPAR

Exponential Amplification Reaction

IA

Isothermal Amplification

LAMP

Loop-Mediated Isothermal Amplification

POCT

Point-Of-Care Testing

qPCR

quantitative Polymerase Chain Reaction

RCA

Rolling Circle Amplification

SDA

Strand Displacement Amplification

AUTHORS’ CONTRIBUTIONS

The authors confirm their contributions to the paper as follows: YG contributed to conceptualization and visualization, YZ assisted with visualization, XL was responsible for data curation, and ZZ, XL, JZ, and YG prepared the draft manuscript. All authors reviewed the results and approved the final version of the manuscript.

CONSENT FOR PUBLICATION

Not applicable.

FUNDING

This work was supported by the Chengdu Medical Research Grant (No. 2022285).

CONFLICT OF INTEREST

The authors declare no conflict of interest, financial or otherwise.

AI DISCLOSURE STATEMENT

During the preparation of this manuscript, the authors used Kimi for grammar improvement and language editing of the Introduction and Discussion sections. After its use, the authors thoroughly reviewed, verified, and revised all AI-assisted content to ensure accuracy and originality. The authors take full responsibility for the integrity and final content of the published article.